Energy Aware Early Detection

Goal

Implementation of Energy Aware Early Detection protocol and testing it with real application.

Introduction

This research is based on initial ideas presented first time in PIMRC’06 conference paper [2]. A comprehensive description can be found in NC54326 [3]. 
The goal of the work is to implement and test a packet classification and queuing algorithm, called Energy Aware Early Detection (EAED). The algorithm is targeted for routing purposes in wireless ad-hoc or mesh networks. The idea of the algorithm in a nutshell is that selected packets from the queue are dropped based on the packet type, length of the packet queue and remaining battery capacity of the wireless router. The usage of algorithm will result an energy savings of the devices and thus increase the lifetime of the network through longer network connectivity.

Initial protocol research description and implementation notes can be found in [1,4]. In this project, EAED protocol will be validated with real application. This means implementing and testing of EAED with some real application (voice or video transmission; e. g. linphone) for simple 3-node scenario (source-dropper-dest). N810 is considered as target device for packet dropper node. One of the tasks here is formulation of criteria for EAED performance evaluation. Such criteria should take into account quality measures (i.e. quality of reconstructed signal) and network lifetime.

Project description

1. Obtaining initial knowledge about Linux kernel development 
   Subtasks: studying scratch box, studying kernel modules, studying maemo kernel compiling and reflashing
   Duration: 0.5 month
2. Development of EAED packer dropper 
   Duration: 1 month
3. Development of battery management routines 
   Duration: 1 month
   Milestone 1
   Deliverables: packet dropper source code, battery routines source code
4. Creation of test-bed 
   Subtasks: choose application for testing, study usage of packet priority in chosen application, creation of test-bed (PC1-maemo-PC2)
   Duration: 1.5 month
5. Testing of EAED using test-bed 
   Subtasks: running experiments, results analysis, investigation of improvements
   Duration: 1 month
6. Preparing final report 
   Duration: 1 month
   Milestone 2
   Deliverables: final reports describing experiments results and analysis
   Project duration: 6 months
   Project start date: 17.08.2009
7. Long-term plan 
   There are several tasks [3], which could be considered after pilot project:
   - Integration with energy aware routing. 
   Testing EAED for more complex scenarios, i.e. many nodes and routing protocol. From our point of view, the topic here is integration of EAED with some energy aware routing protocol. For example, EAED informs routing that battery is low and it will drop packets in near future. 
   - Generalization of the approach for several real-time streams of different priorities. 
   Some current video compression standards (e. g. H.264 SVC) support coding into several streams. These streams have different priorities and provide different quality levels. Currently EAED does not support several real-time streams with different priorities. The goal here is to generalize the EAED for such traffic sources.

Project members

Supervisor from Nokia: Harri Paloheimo
Supervisor from SUAI: Vladimir Bashun
Graduate student: Dmitry Malichenko

References

[1] Harri Paloheimo, Johanna Nieminen, Energy Aware Early Detection Research Plan

[2] Harri Paloheimo, Jukka Manner, Johanna Nieminen and Antti Ylä-Jääski, Challenges in Packet Scheduling in 4G Wireless Networks, IEEE international Symposium on Personal, Indoor and Mobile Radio Communications, Helsinki, 2006

[3] NC54326: Energy Aware Early Detection

[4] Laura Takkinen, Implementation Notes for Energy Aware Early Detection (EAED)